Looking at the internals of the Loras College Statewide Iowa Survey, it seems like it’s a pretty well put together poll. Comparing their sample to the latest Iowa registration numbers, it appears that Loras may have undersampled the NO Party/Independents, but they have about the right mix of Democrats and Republicans. And, in any case, an undersampling of independents probably skews the results towards Donald Trump and Chuck Grassley.
For example, Grassley has a net favorable/unfavorable rating of 48%/42% with independents, and is losing with them to Patty Judge by a 41.9%-48.0% margin. So, if you add more independents to the sample, he probably loses his overall 46% to 45% lead.
Likewise, Donald Trump is getting crushed 48%-34% (-14) in the poll, but among independents he’s losing by a mammoth 44.7%-23.5% (-21) margin.
The survey is made up of 35.0% Republicans, 33.0% Democrats, and 29.8% Independents, but according to the Secretary of State, there are now more independents (670,068) than Republicans (639,476) or Democrats (610,608) who are registered to vote in the Hawkeye State. Maybe the independents don’t turn out at the same rate as party members, so it’s possible that the sample is dead-on. What’s doubtful is that it is skewed toward the Democrats.
Either way, it shows that Donald Trump is not competitive and Chuck Grassley is in a dead-heat. It’d be tempting to blame Grassley’s woes on Trump’s unpopularity (54.7% of Iowans have a very unfavorable view of Trump, and 68.9% have an overall unfavorable view of him), but we know that Grassley has been in the news as the lead architect of the Senate’s refusal to hold a hearing for Merrick Garland, the president’s nominee to replace Antonin Scalia on the Supreme Court. It’s costing him because he’s historically been very popular but he now has a 41.4% unfavorable rating. That’s not terrible, but it’s far below where he’s been in the past. If he’s going to hold on, he’s going to need a lot of crossover votes, but less than a quarter of Democrats (24.8%) have a favorable view of him right now.
Now, Grassley has been in Congress since 1975 and a senator since 1981. He’ll be 83 years old on Election Day. I don’t know if this is really how he wants to go out. I am not even sure why he wants to continue in the job. He’s got to be frustrated. Just this week he had to announce that he almost definitely won’t be able to get his criminal justice reform bill through the Senate this year.
“I don’t see how it gets done before” July 15, Grassley said, referencing the day the senators depart from Washington and won’t return until after Labor Day. “It’s a real big disappointment to me because we’ve worked so hard to do what the leadership wanted to get out more Republican sponsors.”
The criminal justice reform bill was probably the best chance this Congress had to pass a meaningful bill and they can’t get it done. Unless Grassley just likes the prestige and lifestyle of being a senator, I see no reason for him to want to continue. He used to be a legislator, but no Republican can be a legislator in this day and age, and certainly not under a prospective President Hillary Clinton.
If I were him, I’d drop out before Patty Judge cleans his clock in November.
I think these old farts like Grassley just get addicted to the power and prestige, as well as the lobbyist grift. They make way more than their salaries, sothey hang on and on.
The GOP is such a sucking of drain of do-nothing uselessness anymore, it’s gotta be hard to get up every day and pretend to be useful. But the benefits are super!
Much as I had little regard for Tan Man Boehner, his leave-taking was instructive. He’d finally had enough and couldn’t wait to get out of there.
It’s too bad things are so toxic and dysfunctional and show little sign of any improvement.
I think you need to be cautious about any particular poll as polling seems even wilder than usual this year. The shift away from landlines makes polling substantially more difficult technically, and between Trump and the effect of a presidential selection year, turnout models are going to be wobbly too. I’m particularly suspicious about this poll as it’s such a big change from several prior polls, which have shown Clinton with a small lead and Judge distinctly behind.
But I’ll be very happy if it’s true!
Have been hearing the “shift from landlines” excuse for why polling isn’t credible for more than a dozen years. And yet the polling over that period of time has been as accurate as it ever was.
Reporters and the public aren’t educated consumers of polls. They can’t discriminated between good enough and crap polls. Don’t properly weight trends. Don’t understand MOE and probabilities.
This a thousand times.
I recall hearing that Kerry voters in 2004 were being underpolled because of cell-phone-only voters trend younger and more non-white than the population as a whole.
But the 2004 polls were generally right. Those invoking the “cell phone only” hypothesis to discredit polls were the original un-skew-ers. I learned my lesson well then.
When I saw the Romney folks unskewing in 2012, I knew it was the same deal: wishful thinking. Yes, there were a few spectacular polling failures (Gallup, anyone?), but in generally the polls were dead on again.
If you’re going to discredit the polls, the burden is very heavily on the person claiming the polls are wrong. They need real, substantive reasons for believing that. And simply dissecting each poll that you disagree with for a single crosstab that you don’t think is quite right doesn’t cut it.
Not sure why those with no experience or expertise in polling and statistics concluded that the pollsters are too dumb to recognize population and tech changes and incorporate that in their polling methods.
Not enough people around that remember the mega-flub in 1948, and no halfway decent pollster has replicated that error since then. Maybe people generalized from the alleged FL 2000 exit polling flub. It wasn’t a flub but that’s what has been accepted.
Aggregation and trends (which is what Nate Silver does) works better than single point in time polls. However, that gives too much weight to pollsters that are consistently biased. i.e. Gall-up skews GOP. Then there’s the science and art of interpreting the poll results.
Loras was created in 2014. They had the Democratic winning in 2014 by 1 in the Senate race. The Democrat lost by 8. Three weeks out from the Iowa Caucus they had Clinton leading by 29.
So yea, I got my doubts. They have Clinton up by 14, PPP the only other recent pollster had Clinton up 2.
Looks like junk to me, from a pollster with a terrible record.
Garbage in. Garbage out.
But yea – lets not check to see if there are other polls from Iowa to see if this result makes sense.
that’s why I jumped into the internals. They look good.
I find it difficult to take the analysis seriously when you didn’t even bother to compare them against the only other pollster who has been in the field, and that has completely differently results.
There are other reasons to think this is an outlier.
Right now there have been 36 polls taken in June. The average swing from 2012 in those polls 3.07, which predict a margin of 6.9 for Clinton nationally. This method has proven more accurate in predicting the final result since 1996, sometimes significantly so (most notably in 1996, 2000 and 2012). But it also allows the identification of outliers.
Without exception the larger swings are understood: Utah for example. But this poll has no other poll showing a similar shift from 2012 in the purple states.
It does not in any way pass the smell test.
The state polling analysis is here
https://docs.google.com/spreadsheets/d/1IWl01Ik9A2EyaRAgrd0jFBUPMjZS7QCoUYZyhd7C4ns/edit#gid=7153511
63
Yeah, that fooled me a couple of years ago (and I’ve actually taken numerous stat courses). I don’t know what Loras is doing wrong (could be sampling errors, bias, incompetence), but they are the worst.
Marie3–A few weeks ago, when Steven D was promoting a claim that Hillary Clinton was the beneficiary of nationwide electoral fraud, I wrote several comments criticizing the statistical analysis behind the claim. In one of those comments, I wrote something about having experience with statistics in my work as a scientific researcher. You then took me to task for waving my credentials in people’s faces.
And now, here you are waving your credentials in people’s faces. (“I’ve actually taken numerous stat courses.”)
I think it’s great that you’ve taken those statistics courses. But you know what? What’s good for the goose is good for the gander.
If you’re going to challenge somebody, read and comprehend what they’ve written. I know that’s a lot to ask because challenging me has become a kneejerk response from you.
I only mentioned my education in statistics because two years ago and absent any track record on Loras, it led me to that they were doing high quality polling and therefore deserved more weight than the others. The actual results that year and subsequently have demonstrated that Loras isn’t at all reliable. IOW, my education was not only not a help but a hindrance.
You know the statistical difference between sampling “how are you going to vote” and investigating “how did they vote” in don’t you?
I like your confidence, Boo.
Even if I don’t share it.
Loras produces polling reports that are practically textbook perfect. All the i’s dotted and t’s crossed. As such I gave a lot of weight to the first Loras poll I saw (iirc September 2014).
The problem is that the Loras polls suck. No external or actual results validity. IOW — so not credible that they aren’t worth looking at.
Meaning the analysis is carefully done, but the inputs are nonsense?
they could ask the questions in a way that creates bias
For example, Grassley has a net favorable/unfavorable rating of 48%/42% with independents, and is losing with them to Patty Judge by a 41.9%-48.0% margin.
Has Iowa had their primaries yet? is Judge in fact the Democratic nominee?
Never mind. I see that they took place at the beginning of June. What was the caucus turnout in January like, approximately? Because the turnout for the primary seems on the low side.
Getting rid of Grassley would be sublime.
Seeing Grassley leave the Senate would be nice, but there is no way he leaves on his own. He knows that his party is going to lose some seats this cycle, so he will try to hang on to his until the bitter end.